Telegram interface for planning and managing lalph tasks with Claude Code. Draft specs, review architecture diagrams, and approve plans — all from your phone. Also includes PR notifications, auto-merge, and task tracking as extras.
npm i -g @qotaq/lalphgramOr run directly:
npx @qotaq/lalphgram- lalph set up in your project
- A Telegram bot token (you'll be prompted on first run)
lalphgram [options]| Flag | Default | Description |
|---|---|---|
--interval <seconds> |
30 |
Poll interval for GitHub/Linear |
--keyword <word> |
"urgent" |
Trigger keyword for comment timer |
--timer <seconds> |
300 |
Comment timer delay |
On first launch, the CLI will:
- Prompt for your Telegram bot token (stored in
~/.lalph/config/notify.telegram) - Ask whether to enable auto-merge and configure wait time
- Ask you to send a message to your bot to confirm the chat ID
After setup, the event loop starts automatically.
Manage Claude coding plans directly from Telegram:
- Tap Plan → select project → choose plan type (Feature/Bug/Refactor/Other)
- Describe the work → tap Done
- Claude analyzes and produces spec files (architecture diagrams, test plans)
- Review specs via Telegraph link → Approve or Abort
- Ask follow-up questions or interrupt at any point
- New PRs — alerts when PRs are opened
- Merge conflicts — detects conflicts and posts a GitHub comment + Telegram alert
- Broken CI — when CI checks fail on a PR, posts a GitHub comment listing the failed checks, moves the linked issue to "Todo" with urgent priority, and sends a Telegram alert
- New tasks — notifies on Linear/GitHub issue creation and state changes
lalphgram posts comments on PRs automatically, prefixed with [Automatic]:
[Automatic] This PR has merge conflicts that need to be resolved.— posted when a conflict is detected[Automatic] CI checks failed for this PR: ...— posted when CI checks fail, listing the failed check names
Both of these also move the linked issue back to "Todo" with urgent priority. This is designed for use with lalph — when lalph finishes a task it moves the issue to "in-review", but if CI fails or conflicts arise, lalphgram moves it back to "Todo" so lalph picks it up again and fixes the problem.
The [Automatic] prefix distinguishes bot comments from human comments, which matters for auto-merge (see below).
Monitors PR comments and moves linked issues back to "Todo":
- Keyword trigger — if a comment contains the trigger keyword (default:
"urgent"), immediately moves the issue to "Todo" with urgent priority - Timer trigger — any other comment starts a countdown (default: 300s). If no new comment arrives before it expires, the issue moves to "Todo" with urgent priority. Each new comment resets the timer
When enabled, monitors PRs and merges them automatically once:
- All CI checks pass
- A configurable cooldown period has elapsed since the last push
- No manual (human) comments exist on the PR — any comment without the
[Automatic]prefix disables auto-merge for that PR, since it indicates a human is engaged in review
MIT